RAINFALL OVER FOREST REGIONS - INTERCEPTION LOSS

Transcription

1 CHAPTER 7 RAINFALL OVER FOREST REGIONS - INTERCEPTION LOSS

2 " 1 GENERAL Interception is the portion of the precipitation falling on the earth's surface which is stored or collected by vegetal cover and subsequently evaporated. The volume of water so lost is called interception loss. In studies of major storm events and floods the interception loss is generally neglected. But in water balance studies it is a very significant parameter. Precipitation falling on vegetation may be retained on leaves or blades of grass, flow down the stern of plants and become 'stern flow', or fall off the leaves to become part ot the 'through fall'. The amount of water intercepted is a function of (i) the storm character; (ii) the species, age and density of plants and trees; and (iii) the season ot the year. A number of studies on interception have been reported from different parts of the world. Studies on rainfall interception in the forests of New Zealand have been reported by Miller (1963) and Aldridge and Jackson (1973). Rowe (1975, 1979) has reported studies in the mountain beech on the Craigieburn Range of Canterbury and also tor a mixed beech hardwood forest near Reef ton. Dolman (198'/) has compared summer and winter rainfall interception in an Oak forest and tried to predict with an analytical and a numerical simulation model. Studies conducted in Britain 186

3 (Calder, 1976; 1978) reveal that in some climates interception losses can be double that ot the water used by transpiration. Yadav and Mishra (l98~) have given an idea about the distribution of precipitation under a tropical dry deciduous forest stand of central India. The present Chapter describes the interterception study carried out during the southwest monsoon months (June - September) of 1986 in the dense forest at Peruvannamuzhi, Kozhikode District. 7.2 EXPERIMENTAL SITE The site chosen for the study is a small deciduous forest catchment at Peruvannamuzhi, 75 km northeast of Kozhikode city. It lies on the Western Ghats at 11 3~' Nand 75 29' E (Pig 7.1). The average altitude above mean sea level is 350 m having a catchment slope of 9.76 per centage. Average annual rainfall of the region is about 4500 mm, ot which 70% is during southwest monsoon (June-September), 20% during northeast monsoon (October-December) and the remaining 10% distributed during the rest of the period. The mixed forest of the study area comprises about 11 tree species. The height of the trees vary from ~ mto 25 m and the average is 11 m. Diameter at breast height (dbh) varies from 5 cm to 100 cm having an average ot 24 cm. 187

4 ' 11 3!S' no 34' ' FIG ~0 IA9 75 I!SO' 75 I!S,' SCALE o *_ ~,,~V) IIASA\R~,\ 76 10' KERALA SCALE o '0 1001C. '-."-,!;o CZN"ANORE'--!...~ARNATAKA o ~ WYNAO.J '-. PERUVANNAMU~ r:...l TAMILNAOU 10 0' LAKSHADWEEP SEA ~ALICUT.! MAUPUII'1!\ P.:.,..../ I l......, ~ i... ';>!' 1 0' 10 Qo 76 10' REFERENCES STATE BOUNDARY RIVER STREAMS CATCHMENT (STUI7\' I IIOAD contours ~ " ", "'-- r~) ~ INTERCEPTION STUDIES - EXPERIr1ENT..u. SITE AT PERUVA!1N.\MUZHI 75 14" I!lO' 75 I SI', I i (

5 varies trom ~ cm to 100 cm having an average ot 24 cm. Canopy thickness of the region varies from 1.5 m to 11.0 m. In the 5-15 cm dbh class, the average tree density is 275 trees/ha, cm class it is 125 trees/ha; cm class 75 trees/ha and cm class 33 trees/ha. During the study period crown cover of the study area was visually estimated to be about percentage opaque to direct sunlight. The important species in the study area are Dalbergia latifolia, Elaeocarpus serratus, Acacia ferruginea, Aporusa lindleyana, Machilus macrantha, Euodia 1unuakenda, Terminalia paniculata, Diospyros candolleana, Caloplyllum apetalum, Grewia tiliefolia, Ficus arnottiana, Terminalia bellirica, Ailanthus triphysa, Sehleichera oleosa, Bombax insigne, Strychnos nux-vomica, Tectona grandis, etc. Ground flora is rich in species and mostly consisted of shrubs and grasses. Since the study area is a mixed thick forest the branches of the trees are overlapped each other and most of the cases the canopy thickness is not of a single tree but a mixture of different trees. 7.3 EXPERIMENTAL METHODS Incident Daily Rainfall: No attempt was made to measure raintall immediately above the forest crown. Standard ordinary and automatic 189

6 recording raingauges installed at the ground level just outside the forest edge are used to estimate the direct incident raintall. Records trom the six raingauges installed on the boundary of the forest Ghow that there may not be much ditference between the raintall inside the forest and just outside the forest edge Throughfall : Metal Throughfall troughs (Wright, 1977) of size 200 cm x 20 cm x 30 cm (Fig 7.2) were used to measure the throughtall. They were constructed of pre-fabricated mild steel sheet. The troughs were placed, 5 to 10 at a spot interconnected with each other, below the crown at the ground level. The troughs were tilted tor rapid transfer of the water to the measurement device. A shallow tilt increases the response time of the gauge and also increases evaporation loss. A steep tilt menas that the shallow is fast enough to wash any trash towards the drain hole. tlow The water collected through the troughs were measured and compared, in depth, with the readings in the raingauges kept just outside the forest. The troughs were kept as clean as possible to minimise evaporation loss from damp trash in the trough. Ten troughs were used tor keeping randomly under different tree crowns. Daily average values were computed and compared with the incident rainfall. 190

7 \0 r- 10.swa /11 $ $"CCT I I -:.-..,.,:,.; ;VJ v. STEMFlOW GAUGE TROUGH AND COLLAR TYPE STEM FLOW GAUGE I/o,rClllfAlQc,.TIf 0'$/11 ($oc., IlfTCIINAtlCIf.T" 1/11 INrCI/lfAl "/or" 0 '1/11 (IOC,., "XC... 1/$ FtAT ~---, \\ I ",, " ----,,---- ~/.Xl.)($... 1/$ A"alC_ ---, I I -----{ ", " -', I " Mei' Al THROUGHFALL TROUGH -{---- " FIG 7.2 THROUGHFALL '-

8 Since the position ot the troughs were changed frequently, it was observed that on an average the excess readings obtained due to 'drip points' had neutralised with the 'null points'. During large storms water leave branches at some bends and low points in a steady trickle; it is meant by 'drip point'. 'Null Point' is where there are no drop even during large storms Stemflow: Collar type gauge has been used to compute the stemflm.r. Thick (1 mm) and stift polythene collars were sealed to the trunks with bitumastic sealing compound and had drain-tubes to conduct the water to the measuring device. The crown areas were estimated from the vertical projection of the edge of the crown to the ground and used in computing the net stemf10w. In the present study six trees were selected, of which the crown areas ranging from sq.m to sq.m. and dbh from 12.1 cm to 51.6 cm. Th~ height of the trees selected for the study ranges from 5 m to 22 m. Stemflow varies mainly with dbh, crown area, canopy thickness. Therefore the trees for stemtlow studies were selected in such a way that the average ot each parameter the selected trees are almost coinciding with the average~ ot of the total trees in the study area. 192

9 During the southwest monsoon period, most ot the days the raintall is either continuous or with very less time-gap between showers in the Western Ghats region ot Kerala. Hence, measurements were made daily during June, July, August and September of To minimise the evaporation losses from the measuring devices, polythene covers were used on the top of the collecting jars and a thin layer of coconut oil was applied daily over the water content. By these methods it was found that the evaporation loss was almost zero from the measuring devices of the gauges. During heavy raining, intermittent measurements were made to prevent overflow from the collectors and in most of the cases data was collected immediately atter very mild showers. General views of throughtall troughs and stemtlow gauges are given in Figs 7.3 and 7.4, respectively. 7.4 RESULTS Throughta11 The values in Table 7.1 indicate that during the monsoon period throughtall did not occur when the daily rainfall is less than 2.0 mm. In the range ot 2 to 2.9 mm raintall, only an average ot 6% ot the rainfall occurred as throughfall under the thick mixed forest, but it increased with the rainfall. In the case of 4 to 4.9 mm range of 193

10 .. ~ FIG _7. 3 '" THROUGHFAll MEASUREMENT FIG-7. 4 STEMFlOW MEASUREMENT ; ; '.,., 1<-,., > ';.

11 Table r/.l: Distribution ot daily incident rainfall under the forest stand as throughtall, stemflow and interception loss Date of Incid- Average total (mm) % of incident rainfall measure- ent ment daily Thro- Stem Interc- Thro- Stem Intercrain- ugh flow eption ugh flow eption tall fall loss fall loss (mm) '1.5 8' ' '/ ' '

12 Table 1.1 (contd.) Date at Incid- Average total (mm) % at incident raintall measure- ent ment daily Thro- Stem Interc- Thro- Stem Intercrain- ugh flow eption ugh flow eption fall fall loss tall loss (mm) ' '7.' " " " '/ '/ '/ '/ '/ '/

13 Table 1.1 (contd.) Date ot measurement ' Incident daily rainfall (mm) Average total (mm) Through fall Stem flow Interception loss % ot incident rainfall Through tall ' Stem flow Interception loss

14 raintall, the average throughfall was 36% and in the range of l~ to 19.9 mm it was 75%. When the rainfall exceeded 50 mm there was no much steep change in the through throughtall percentage (Fig. 7.5). Linear regression analysis was used to determine relationship between daily raintall and throughfall. In Fig. 7.6 the throughtall is shown in relation to daily rainfall. The linear regression ot mean throughfall (T mm) on daily raintall (R mm) using the data tor 79 days gave the equation. T = R Daily rainfall versus throughfall is plotted in Fig. 7.7 and the regression line is drawn. It can be seen that throughtall linearly increases with the daily raintall and has a strong relationship with each other. The correlation coetticient between throughtall and daily raintall is compl1ted as r = , which is almost a perfect correlation and is highly significant. Separate regression analysis were done tor difterent ranges of raintall. The regression equations and the correlation coefficients are given in Table 7.2 for different ranges of rainfall. All the equations give almost identical results when used to estimate the mean throughfall trom daily raintall. 198

15

16 o 10...J...J ~ Z <t 0:: 7 NET R A I N FA L L THROUGH FALL '.' STEMFLOW o DAILY RAINFALL (mm) Fig 7.6STEMFLOW, THROUGHFALL AND NET RAINFALL AS PERCENTAGE OF DAILY RAINFALL N

17 N o... T = R o DAILY RAINFALL (mm) Rg 7.7 RELATION OF THROUGHFALL (T) TO DAILY RAINFALL (R)

18 Table 7.2: Regression equations and correlation coetficients between throughtall and raintall Ranges of rainfall(mm) Regression equation (T) Correlation coeff. er) o R - 1.!> R R Stemflow As in the case of throughtall, average stemtlow increased with the increase ot raintall. In the range ot 2 to 2.9 mm of raintall the average stemtlow was less than one percentage. In some cases it was very little, not even measurable. In Fig 7.6 the increase in stemflow with the raintall is shown. From the data collected, it was noticed that tor a rainfall of mm, the average stemtlow was only mm. The characteristics ot the trees selected for the stemtlow studies are shown in Table 7.3 and the variations in average stemtlow in ditferent trees with difterent ranges of raintall are illustrated in Table

19 Table 7.3 Characteristics ot the torest trees selected tor stemtlow recording SI No. Species '* Height dbh Crown Crown Crown (m) (cm) area thickness coverage (sq.m) (m) (%) '* 1. Aporusa lindleyana 2. Machilus macrantha 3. Euodia lunuakenda 4. Terminalia paniculata 5. Terminalia paniculata 6. Ailanthus triphysa!:) !) 7!) ' '* Crown thickness and coverage including branches of nearby tress 203

20 Table 7.4 Variations in average stemtlow (%) in the selected trees (as shown in'rable : ".3) with ditterent ranges of raintall Range ot raintall (mm) Trees selected for stemflow measurements Linear regression analysis using the data ot the monsoon months, in which rainfall varies from 0 to mm,, gave the following relation between the average depth of stemflow (S) and daily raintall (R). -2 S = x 10 R In Fig. 7.8, stemtlow versus daily rainfall is plotted and the regression line is drawn. Between 20 and

21 N o U E 3: 6- o...j u.. :::E 5 w I- Cl) x x _ S = x 10. R o DAILY RAINFALL (m m) Fig 7.8 RELATION OF STEMFLOW (5) TO DAILY RAINFALL (R)

22 mm ot raintall the stemtlow showed a slight decrease compared to the line ot best tit. The correlation coettici.ent between stem flow and the daily raintall is tound to be , which is signiticant. The regression e~lations and the correlation coetticients between stemtlow and daily rainfall tor difterent ranges are illustrated in Table 7.5. Table 7.5 Regression equation and correlation coefticient coetficients between stemtlow and raintall Ranges ot raintall(mm) Regression equation (S) Correlation coeff.(r) o R R R Net Raintall Net rainfall is the water that reaches the surface ot the ground as throughfall and stemflow. The sum ot mean throughtall (T) and the mean stemtlow (S) was llsed as the net rainfall (N). That is the ditterence between the incident raintall and the lost raintall due to interception. Since N is dependent of T and S, naturally when T and S increase N also increases. 206

23 The percentage ot lost raintall is less in the range ot 100 to mm, when compared to 0 to 19.9 mm rainfall. But when it was considered in quantity-wise the lost rainfall was high in the tirst case. Specitically, in the range of 2 to 2.9 mm raintall, the average lost rainfall was 93.21% of incident rainfall and in mm range, it was 4.68%. Depth-wise it was 2.12 mm and 5.7 mm, respectively. Linear regression analysis using the net rainfall and the daily raintall, varies from 0 to mm, gave the following relation between the net rainfall (N) and daily raintall (R). N = R Using the above equation the line of best fit is obtained and is ploteed in Fig The correlation coefficient computed between net rainfall and daily rainfall show a value ot r = , which is positive and highly significant. Separate regression analysis were also done tor ditferent ranges of rainfall. The regression equations and the correlation coefficients are given in Table 7.6 for different ranges of rainfall. 207

24 N o (Xl o DAILY RAINFALL (mm) Fig 7.9 RELATION OF NET RAINFALL (N) TO DAILY RAINFALL (R)

25 Table 1.6 Regression equations and correlation coetticients between net raintall and daily raintall Ranges of rainfall(mm) Regression equation (N) Correlation coett.(r) o R R R The results based on the data of southwest monsoon months ot 1986 reveal that 10.45% of the total incident rainfall has been lost as inteception. During the period, throughfall is 84.67% and stemflow is 4.88%. In terms of depth correlation between throughtall and interception loss is positive but it is negative between percentage throughfall and percetnage interception loss which indicates that the percentage interception loss decreases with the increasing proportion of throughfall and also with stemflow and amount of incident precipitation. Because of the retention ot a good amount of rain water in the forest canopy, it was noticed that the humidity of the region is very high compared to the open area. This will prevent evaporation from forest soil. Because of the good coverage of green leaves, the gravitational torce ot 209

26 rain drops are minimised on the way, which is a natural prevention against soil erosion. The study reported is only a footstep of novel studies to be carried out in the torest region of Western Ghats. The study has to be continued in the months other than southwest monsoon period and it should be extended to different forest types. Since Kerala is covered by about 20% of forest, similar type of studies have got vital importance while considering the water balance of the region. 210